Buried lateral index guided lasers and lasers with lateral current blocking layers

CROSS-REFERENCE TO RELATED APPLICATIONS

0001This application is related to U.S. patent application Ser. No. 11/300,861 entitled “High Power Semiconductor Device with Low-Absorptive Facet Window”, and U.S. patent application Ser. No. 11/300,962 entitled “A system for adjusting the wavelength light output of a Semiconductor Device using hydrogenation”, and U.S. patent application Ser. No. 11/300,871 entitled “On-Chip Integration of Passive and Active Optical Components Enabled by Hydrogenation”, all assigned to the same assignee and filed on the same day on Dec. 15, 2005, and all are hereby incorporated by reference.

BACKGROUND

0002Fabricating a semiconductor laser often involves tight confinement of carriers and waveguiding of light generated in an active region of the semiconductor. Vertical confinement and index guiding may be easily achieved by the deposition of cladding layers with the desired index of refraction, both below and above an active region of the semiconductor laser. However, lateral index guiding has proven to be more of a challenge.

0003In order to provide strong lateral index guiding of a laser, a higher bandgap material typically is formed laterally adjacent to the active region. In order to achieve the higher bandgap laterally adjacent regions, traditional fabrication techniques involve etching a ridge next to the active region. A higher bandgap epitaxial material is then regrown beside the ridge as described in Heterostructure Lasers Part B, H. C. Casey and M. B. Panish, Academic Press, Inc., 1978, pages 213-215, ISBN 0-12-163102-8. This etching and regrowth process is undesirable because of the high cost associated with pre-regrowth sample preparation, the epitaxial regrowth process itself, and the manufacturing logistics involved. An additional drawback is the non-planar morphology that results.

0004Thus a simpler method of creating a transverse index guiding mechanism that does not involve etching followed by regrowth is needed.

SUMMARY

0005A semiconductor structure is described. The structure includes an active region. Laterally adjacent to the active region is a hydrogen induced bandgap shifted InGaAsN material formed alongside the active region positioned to improve lateral carrier confinement. In one embodiment, a selected lateral hydrogen profile generates a lateral index variation. The lateral index variation creates a desired optical mode profile.

BRIEF DESCRIPTION OF THE DRAWINGS

0006<figref idref="DRAWINGS">FIG. 1</figref> shows a graph plotting the output photoluminescence spectra of InGaAsN/GaAs quantum wells doped with various quantities of hydrogen.

0007<figref idref="DRAWINGS">FIG. 2</figref> shows a plot of the absorption characteristics of a semiconductor as a function of the incident photon energy.

0008<figref idref="DRAWINGS">FIG. 3</figref> shows a cross sectional view of a laser structure that relies on hydrogenated facets to minimize absorption.

0009<figref idref="DRAWINGS">FIG. 4</figref> shows a mask with apertures being used to control hydrogenation of a wafer.

0010<figref idref="DRAWINGS">FIG. 5</figref> shows a buried index guided laser diode structure using hydrogenated InGaAsN and GaAsN layers for lateral index guiding.

0011<figref idref="DRAWINGS">FIG. 6</figref> is a table showing different confinement factors and effective refractive indexes for various example structures.

0012<figref idref="DRAWINGS">FIG. 7</figref> is a plot of lateral confinement factors for a buried lateral index guided laser structure versus waveguide width for different refractive index steps.

0013<figref idref="DRAWINGS">FIG. 8</figref> shows an example of a ridge-waveguide laser diode.

0014<figref idref="DRAWINGS">FIG. 9</figref> is a schematic that shows different contributions to the net optical gain of a laser as a function of wavelength.

0015<figref idref="DRAWINGS">FIG. 10</figref> shows an array of lasers coupled to gratings, each laser outputs a different frequency of light, all lasers may be on the same wafer.

0016<figref idref="DRAWINGS">FIG. 11</figref> shows a wafer in a vacuum chamber as one method of selectively hydrogenating regions of a wafer.

0017<figref idref="DRAWINGS">FIG. 12</figref> shows a half tone mask using different aperture densities to control hydrogenation of a wafer.

0018<figref idref="DRAWINGS">FIG. 13</figref> shows a VCSEL where the hydrogen content of the active region may be adjusted to tune the frequency of the laser output.

0019<figref idref="DRAWINGS">FIG. 14</figref> shows a cross sectional view of an index guided optical waveguide where a half tone mask is used to create a desired lateral index variation.

0020<figref idref="DRAWINGS">FIG. 15A</figref> is a side cross sectional view of the integration of an optical waveguide with a laser diode. A half tone mask is used in this case to gradually modify the lateral and vertical index profile. <figref idref="DRAWINGS">FIG. 15B</figref> shows the index profile at two sample locations in the structure of <figref idref="DRAWINGS">FIG. 15A</figref>.

0021<figref idref="DRAWINGS">FIG. 16</figref> shows a three dimensional view of a various regions of a semiconductor waveguide coupling a glass fiber to a semiconductor laser.

0022<figref idref="DRAWINGS">FIG. 17</figref> is a cross section view of a two section laser diode structure with a modulator and amplifier region.

0023<figref idref="DRAWINGS">FIGS. 18-21</figref> show various integrated optical circuits that may be fabricated on a single semiconductor wafer using the described waveguides.

0024<figref idref="DRAWINGS">FIG. 22</figref> shows a plot of an asymmetric index profile plotted along a lateral or vertical dimension and how the index profile affects the optical mode.

0025<figref idref="DRAWINGS">FIG. 23</figref> shows a plot of a symmetric index profile that may be used, for example, in a beam splitter and how the index profile affects the optical mode.

DETAILED DESCRIPTION

0026A system for forming and interconnecting optoelectronic devices is described. The system controls exposes a semiconductor compound including at least GaAsN to hydrogen in a hydrogen concentration adjustment process. More typically, the GaAsN is an alloy that includes at least one of either aluminum to form AlGaAsN or Indium to form InGaAsN. For convenience, the specification will refer to InGaAsN, although it should be understood that the invention should apply to any semiconductor material that changes bandgap as a result of exposure to hydrogen, and the most common semiconductor materials that exhibit this property are alloys, typically AlGaAsN or InGaAsN. By controlling the amount of hydrogen in the sample, the bandgap of the sample can be controlled thereby controlling the sample's refractive index as well as absorption properties.

0027As used herein, “altering” or “changing the hydrogen concentration” is broadly defined to include all methods of increasing or decreasing the concentration of hydrogen in a material. The adjustments may be accomplished by either removing previously incorporated hydrogen or by adding hydrogen atoms and/or molecules. Incorporation or otherwise adding hydrogen to a material may be accomplished using various techniques, including but not limited to, diffusion through exposure of a sample to a hydrogen gas. In one example, a layer is grown by Metal Organic Vapor Phase Epitaxy (MOCVD), where hydride precursor gases such as ammonia are typically employed. Moreover, hydrogen is usually the carrier gas of choice for flowing metal organics for building the epitaxial structure. The hydrogen-rich environment in the entire growth process makes it easy to incorporate hydrogen into the material during epitaxial growth. Masks may be used to control the regions that are exposed to hydrogen.

0028Alternately, changing hydrogen concentrations may refer to a hydrogen ion implantation wherein high energy hydrogen particles are directed at the sample to selectively incorporate hydrogen into a sample. Ion implantation, selective areas of the material can be masked off by photoresist during device fabrication. The unmasked surfaces can then be exposed to standard hydrogen ion implantation accelerated at, typically, between 50 KeV to 350 KeV depending on the penetration depth desired. In practice, the exact dosage and energy levels required is determined experimentally using Secondary Ion Mass Spectroscopy (SIMS) characterization in conjunction with electroluminescence measurements. Monte Carlo computer simulation techniques using software such as the popular SRIM and TRIM packages can also be employed to model the ion implantation process. Sources of hydrogen include, but are not limited to low energy Kaufman sources and plasma hydrogenation systems.

0029As defined and used herein, altering or changing hydrogen concentrations is not limited to increasing hydrogen concentrations. Hydrogen concentration changes also include processes of removing previously incorporated hydrogen atoms from a pre-hydrogenated sample. An example method of reducing hydrogen concentration is to anneal portions of a sample, typically at temperatures about 500 degrees centigrade or above during fabrication. Controlled removal from specific regions may be effected by patterning a region with a “hydrogen getter” such as palladium and then heating the sample. Electron beam irradiation or wavelength selected laser writing also provide methods of localized heating for selective area hydrogen removal.

0030Changing the hydrogen concentrations in InGaAsN materials produces a bandgap shift in InGaAsN materials or samples. The process is described in G. Baldassarri, M. Bissiri, A. Polimeni, and M. Capizzi, “Hydrogen-induced Band gap Tuning of (InGa)(AsN)/GaAs Single Quantum Wells” Appl. Phys. Lett., Vol. 78(22), pp. 3472-3474, (2001) which is hereby incorporated by reference. The effects of hydrogen on the InGaAsN bandgap may be measured by observing the photon energy of the photoluminescence peak of InGaAsN/GaAs quantum well at different hydrogen concentrations.

0031<figref idref="DRAWINGS">FIG. 1</figref> (taken from the above cited Baldassarri reference) plots the output photoluminescence spectrum of a plurality of In_0.34Ga_0.66As_1-yN_y/GaAs quantum well structures as a function of quantum well hydrogen concentrations. The dashed curves <b>104</b>, <b>108</b>, <b>112</b> show the photoluminescence of a reference sample with y=0 (thus no nitrogen) while the solid curves show the photoluminescence of a reference sample with y=0.007. H_Oin the figure represents a hydrogen concentration that results from exposures to a hydrogen flux of 1×10¹⁶hydrogen atoms/cm²

0032In the absence of hydrogenation, the energy difference between the photons output by an InGaAsN active region and the photons output by an InGaAs (no nitrogen) active region is approximately 0.05 electron volts (eV). In particular, curve <b>116</b> peaks around 1.01 eV indicating the majority of photons output by InGaAs has an energy of 1.01 eV. Curve <b>104</b> peaks around 1.06 eV indicating the majority of photons output by InGaAsN have an energy of around 1.06 eV. Increasing hydrogenation levels (increased hydrogen concentrations) increases the output photon energies of nitrogen containing materials. In particular, at a concentration of 690 H_O(a concentration resulting from a flux of 690×10¹⁶hydrogen atoms/cm²) shown in curve <b>120</b>, the bandgap of In_0.34Ga_0.66As_0.993N_0.007/GaAs shifts by about 60 meV resulting in a bandgap similar to that of a nitrogen free reference sample shown by curve <b>112</b>.

0033Changing the bandgap of a material not only alters the light output by a material, it also has other effects. For example, changing the bandgap changes the refractive index of a material and also the photon absorption properties of the material. As the InGaAsN material bandgap increases due to increased hydrogen concentration, the refractive index decreases. A typical change in the index of refraction is an index shift from about 3.68 to an index of about 3.52 for light with a 1.3 micrometer wavelength. The associated refractive index shift is described in T. Kitatani, M. Kondow, K. Hinoda, Y. Yazawa and M. Okai, “Characterization of the Refractive Index of Strained GaInNAs Layers by Spectroscopic Ellipsomtery,” Jpn. J. Appl. Phys. Vol. 37, pp. 753-757, 1998 which is hereby incorporated by reference in its entirety. By controlling the refractive index change in different parts of a device, optical devices and interconnects may be formed as will be described below.

0034<figref idref="DRAWINGS">FIG. 2</figref> plots a material's absorption properties as a function of incident photon energy. The absorption is plotted on axis <b>204</b> and the energy of an incident photon plotted on axis <b>208</b>. Photons below a bandgap energy <b>212</b> are not absorbed and thus the thus the material appears transparent to photons below bandgap energy <b>212</b>. Photons with energies above the bandgap encounter some absorption. Curve <b>216</b> shows that as the photon energy increases, the absorption coefficient also increases. Changing the hydrogen concentration shifts bandgap energy <b>212</b> along axis <b>208</b> resulting in changes to the material absorption properties. As will be described, this property will be used to lower the absorption of the facets in a laser diode.

0035A number of applications can be made using the previously described properties. In one example, the lateral as well as vertical hydrogen concentrations are manipulated to modify the bandgap and refractive index. Step changes in the index of refraction can be used to form waveguide boundaries. However, graded or gradual index variations may also be used to form waveguides or control optical mode profiles.

0036One example of using index changes to control modal profile are shown in <figref idref="DRAWINGS">FIGS. 22 and 23</figref>. <figref idref="DRAWINGS">FIGS. 22 and 23</figref> show a plot of a varied index profile <b>2204</b> and <b>2304</b> plotted along a lateral (or vertical dimension) (axis d). The index profiles affect the optical modes of light propagating through the varied index profile. Curves <b>2208</b> and <b>2308</b> show example resulting optical modes for each index variation curve. More details on the use of both the stepped index variation and the graded index variation will be described in the devices that follow.

0037High Power Semiconductor Light Output Devices with Low-Absorptive Facet Windows

0038The described methods may be used to alter the facets of semiconductor light output devices, although particular use may be found in a traditional high powered long wavelength InGaAsN system laser. As used herein, “semiconductor light output device” is broadly defined to include any device that uses a gain region of a semiconductor material to amplify or to generate light. Example semiconductor devices include lasers, amplifiers. As used herein, “facets” are broadly defined to mean the region immediately adjacent an active region through which laser light passes before exiting the laser. The facet is often made of material very similar to the active region. The “active region” is defined as the area of a semiconductor device that generates and/or amplifies light usually by stimulated or spontaneous emission of light. Thus “active region” is typically the gain medium of a laser, or an amplifier. In a LED (including superluminescent LEDs) the “active region” is usually the region in which spontaneous emission of light occurs. As previously described in the background, laser facets often absorb some of the laser's output energy. The absorbed energy heats the facet causing defects and bandgap shrinkage. The resulting defects and bandgap shrinkage increases facet energy absorption sometimes resulting in catastrophic optical damage and laser failure.

0039In order to avoid the catastrophic damage, one exemplary embodiment of the invention adjusts the concentration of hydrogen in the facet region. In particular, hydrogen content in the InGaAsN and/or GaAsN facet region is increased to increase the bandgap in the facet regions rendering the facet less-absorbing of photons output by a semiconductor device. In particular, facet hydrogenation makes the facet bandgap slightly larger then the bandgap of the active region

0040Various methods may be used to change the hydrogen concentration in the facet. One method of increasing the hydrogen concentration is hydrogen implantation in the facet region. A second method is simply to allow hydrogen to diffuse into the cleaved facet region of the semiconductor laser. <figref idref="DRAWINGS">FIG. 3</figref> shows a laser <b>300</b> exposed to hydrogen <b>304</b> such that the hydrogen diffuses into one or both laser facets <b>308</b>, <b>312</b>. Such hydrogen incorporation may be achieved by cleaving the laser to create laser ends or laser facets. The cleaved facets, typically laser facets <b>308</b> and/or <b>312</b> are then exposed to monatomic hydrogen at about 300 degrees centigrade. The rate of diffusion depends on temperature and can be modeled theoretically as described, for example, in Physics of Semiconductor Devices, 2^ndEdition, S. M. Sze, John Wiley &amp; Sons, 1981, pp. 66-69 and in “Modeling of hydrogen diffusion in p-type GaAs:Zn”, R. Rahbi, et. al., Physica B: Physics of Condensed Matter, Volume 170, Issue 1-4, p. 135-140. The diffusion behavior can be determined experimentally using standard Secondary Ion Mass Spectroscopy (SIMS) characterization techniques. A typical diffusion depth of between about 500 nm to 5 μm is desirable. A typical exposure may occur for a period of time of between 1 and 15,000 minutes in a chamber with hydrogen at a pressure of 10 to 800 torr. Facet HR coating or passivation may occur after exposure and diffusion of the hydrogen into the facet. The hydrogen levels are typically increased until a bandgap shift of between 1 meV and 100 meV is achieved.

0041Sometimes, processing of individual lasers is slow and a wafer level processing is needed. <figref idref="DRAWINGS">FIG. 4</figref> shows one example of masking a wafer <b>400</b>. In wafer level processing, the hydrogenation of the facets may be completed prior to cleaving or sawing the wafer to create individual lasers. In the embodiment shown in <figref idref="DRAWINGS">FIG. 4</figref>, a metal or dielectric mask <b>408</b> covers wafer <b>400</b>. The mask <b>408</b> includes opening <b>412</b> positioned over where the laser will eventually be cleaved. The wafer <b>400</b> is then exposed to hydrogen atoms <b>416</b>. Hydrogen atoms impinging the wafer <b>400</b> through openings <b>412</b> and diffuse into the active layers. In an alternate exemplary embodiment, ion implantation drives hydrogen into the facet area prior to laser facet cleaving. In still a third exemplary embodiment, the entire wafer is hydrogenated and hydrogen then hydrogen removed from non-facet areas. Removal may be accomplished by selective heating of non-facet areas or by masking non-facet areas with a hydrogen attracting material. After the hydrogen is embedded into the laser facets, the wafer may be cut or cleaved to form individual devices such as lasers, LEDs and the like.

0042In regions that include nitrogen, hydrogen increases bandgap in facets <b>308</b>, <b>312</b> such that the facet bandgap is larger than the bandgap in region <b>316</b> between the facets. Thus, InGaAsN layer <b>320</b> includes a facet region <b>324</b> that has a slightly larger bandgap than active region <b>328</b>. Likewise, GaAsN spacing layers <b>332</b> includes facet regions affected by hydrogenation and thus have a bandgap larger than regions that do not contain nitrogen. Examples of regions that do not contain nitrogen include AlGaAs guiding layers <b>336</b> and AlGaAs:Si cladding layers <b>340</b>. A typical example hydrogen concentration in the laser facet <b>308</b>, <b>312</b> region exceeds 5×10¹⁶hydrogen atoms/cm²which results in a bandgap shift that exceeds 20 meV. In one embodiment InGaAsN active region outputs light at 1.48 micrometer which results from a bandgap of about 0.84 eV. Very few photons of this energy are absorbed by the laser facet with a bandgap that typically exceeds 0.86 eV.

0043Besides increasing the bandgap, hydrogen in the facets also reduces the facet index of refraction. A reduced waveguide refractive index differential between the guiding layers <b>336</b> and the facet <b>308</b>, <b>312</b> increases beam spreading at the facet. Beam spreading lowers the beam power density in the facet which also helps to reduce facet heating. In one embodiment, the lateral index of refraction in the facet is graded to provide additional control of the optical modes. The optical modes may be molded to further enhance beam spreading of light from the active region of the semiconductor device. Alternately, the optical mode may be shaped for input into a receiving device or waveguide.

0044High powered long wavelength high reliability lasers such as 1.48 micrometer lasers are particularly useful for optical communication systems. The 1.48 micrometer wavelength which can be generated by an InGaAsN laser, is particularly suitable for pumping erbium-doped fiber amplifiers often used in long haul optical communication systems. Such a communication system is described in, Understanding Fiber Optics, 2^ndEdition, Jeff Hecht, Sams Publishing, 1993, ISBN 0-672-30350-7, pp. 112-113 which is hereby incorporated by reference.

0045Pump wavelength at 1.48 μm allows longer and more uniform fiber amplifier gain sections compared to the alternative 980 nm pump wavelength because of lower fiber absorption at 1.48 μm. Strong absorption of the pump source at 980 nm results in weak pumping of the signal at greater distances from the laser source. Thus, InGaAsN pump lasers at 1.48 μm is desirable, and the use of hydrogenated facets in high power InGaAsN pump lasers allows these lasers to operate with long lifetimes due to minimized facet heating.

0046The signal wavelength in long haul communications systems is typically around 1.55 μm. InGaAsN active layers can be designed for 1.55 μm laser operation. 1.55 μm is the signal wavelength of choice in long haul fiber-optics communication systems because of the wavelength's low absorption in typical fibers. High power at 1.55 μm is advantageous because it allows the signal to travel a longer distance before requiring amplification thereby allowing fewer costly amplifiers in the communication link. Thus the high powered signal lasers that output 1.55 μm also benefit significantly from reduced facet heating. Thus both the pump lasers and the signal lasers operate under high power conditions, the reliability of which can be enhanced by minimal absorption and heating at the laser facet. The increased reliability is particularly valuable in applications involving hard to access areas such as undersea fiber links.

0047Buried Lateral Index Guided Lasers and Lasers with Lateral Current Blocking Layers

0048In order to improve the performance of semiconductor light output devices, usually semiconductor lasers, traverse index guiding mechanisms are used to laterally index guide the active region. In the following description, a hydrogenation-induced bandgap shifted (HIBS) material along at least one edge of an InGaAsN active region serves as the lateral index guiding structure. Such a structure is shown in <figref idref="DRAWINGS">FIG. 5</figref>.

0049<figref idref="DRAWINGS">FIG. 5</figref> shows an InGaAsN square quantum well active region <b>504</b> sandwiched between GaAs (N) spacers <b>508</b>, <b>512</b> and AlGaAs cladding layers <b>516</b>, <b>520</b>, <b>524</b>, <b>528</b>. As used herein “sandwich” merely means between two layers, and not necessarily adjacent to either of the two layers. In the illustrated embodiment, closer cladding layers or guiding layers <b>516</b>, <b>520</b> are undoped while lower cladding layer <b>524</b> is silicon doped and upper cladding is Carbon doped. Substrate <b>532</b> supports the structure while contacts <b>536</b>, <b>540</b> allow electrical pumping of active region <b>504</b>. Such an InGaAsN laser stack is described in many prior art references including U.S. Pat. No. 6,922,426 entitled “Vertical Cavity Surface Emitting Laser Including Indium in the Active Region” by Johnson which is hereby incorporated by reference.

0050Hydrogenated lateral region <b>544</b>, <b>548</b> bordering active region <b>504</b> guides photons generated by active region <b>504</b>. As used herein, “lateral” is defined as regions to the side, typically in a wafer that includes multiple layers, lateral regions to an region in a layer, (such as an active region), will be other regions in the same layer. Typically, the lateral region is made from the same material and layers as the material in other regions of the InGaAsN laser except that the lateral regions have been hydrogenated. In one embodiment, lateral regions are merely hydrogenated regions of a laser layer. Thus, when an InGaAsN active region is created from an epitaxially grown layer, the lateral regions are merely portions of the InGaAsN epitaxially grown layer that has been hydrogenated. Adding hydrogen increases the lateral region bandgap and decreases the index of refraction. Both characteristics enhance index guiding of the wave propagating in active region <b>504</b>.

0051Various methods may be used to change the hydrogen concentration in lateral regions <b>544</b>, <b>548</b>. In one example method, hydrogen ion implantation is used to direct hydrogen to the target area, lateral region <b>544</b>, <b>548</b> in the illustrated example. In a second example method, monatomic hydrogen may be diffused through openings in a masked wafer surface. The masking structure of <figref idref="DRAWINGS">FIG. 4</figref> may be used, except that openings <b>412</b> in the mask correspond to surrounding lateral regions <b>544</b>, <b>548</b> instead of laser facets. Thus, in this second method diffusion occurs in a vertical direction. In still a third method, hydrogen may be allowed to laterally diffuse from an etched ridge into surrounding lateral regions <b>544</b>, <b>548</b>. In particular, a ridge may be etched adjacent to lateral regions <b>544</b>, <b>548</b> and the ridge exposed to a hydrogen gas. Different temperatures and pressures may be used to incorporate the hydrogen, one example of which is described in hydrogen diffusion and acceptor passivation in p-type GaAs from R Rahbi et al. APL 73 pp. 1723-1731 (1993). In the cited reference, hydrogen was diffused in the samples by exposure to a hydrogen rf plasma (13.56 MHz) at a constant pressure of 1.2 mbar. The exposure temperature was chosen in the 50-300 Centigrade range for different durations of 30 min to a few hours.

0052As described earlier, changing the hydrogen concentration is not limited to adding hydrogen to a region. In still a fourth method of hydrogenating the lateral regions, hydrogen may be introduced to the entire layer including active regions and lateral regions. Excessive hydrogen may then be removed, particularly from the active region, by selectively heating the active region or by placing the active region in contact with a material with a strong affinity for hydrogen.

0053<figref idref="DRAWINGS">FIG. 6</figref> and <figref idref="DRAWINGS">FIG. 7</figref> illustrate the resulting confinements that may be obtained using lateral hydrogenation. <figref idref="DRAWINGS">FIG. 6</figref> shows confinement factors (ratio of the modal gain to the active region gain) and effective refractive indexes for example InGaAsN laser structures. Rows <b>604</b> and <b>608</b> of <figref idref="DRAWINGS">FIG. 6</figref> shows effective index changes in which only the InGaAsN active layer lateral regions <b>552</b>, <b>556</b> are hydrogenated. Rows <b>612</b> and <b>616</b> show effective index changes when both the InGaAsN active layer lateral region <b>552</b>, <b>556</b> and the GaAsN spacer lateral region <b>560</b>, <b>564</b> are hydrogenated resulting in an active region that is bounded on all fours sides by hydrogenated material. The sample computations are done for an 8 nm wide InGAsN SQW active region and a 35 nm wide GaAs(N) spacer layer.

0054In <figref idref="DRAWINGS">FIG. 6</figref>, column <b>620</b> provides confinement factors and column <b>624</b> provides effective refractive indexes for a zero order mode for various exemplary structures described in Column <b>628</b>. Row <b>604</b> provides the confinement factor and the effective index of refraction of the central region, including a GaAs spacer, all of which has not undergone hydrogenation. Row <b>608</b> provides the confinement factor and the effective refractive index of the adjacent surrounding lateral region when only the lateral portion of the InGaAsN active region is hydrogenated but the lateral portion of GaAs spacers have not been hydrogenated. Box <b>632</b> shows the difference in effective refractive index produced by hydrogenating only the lateral portion of the InGaAsN active region. The 0.004 effective refractive index difference provides some index guiding, however, a larger change in index would provide stronger index guiding.

0055An order of magnitude increase in the index differential may be achieved by not only hydrogenating the active region, but also hydrogenating the adjacent GaAsN spacer lateral region <b>560</b>, <b>564</b>. Row <b>612</b> provides the confinement factor and the effective index of refraction of a central region that has not undergone hydrogenation. Row <b>616</b> provides the confinement factor and the effective refractive index of the adjacent surrounding lateral region when both the lateral portion of the InGaAsN active region and the lateral portion of GaAsN spacers have been hydrogenated. Box <b>640</b> shows that an approximately 0.042 difference in effective indexes results between the effective index of a non-hydrogenated structure and the effective index of a structure where both the lateral active region and lateral spacers have been hydrogenated.

0056<figref idref="DRAWINGS">FIG. 7</figref> plots percentage lateral confinement along axis <b>704</b> as a function of waveguide width along axis <b>708</b> for various example refractive index differences. The example data is for a laser that outputs 1.3 micrometer wavelength light. <figref idref="DRAWINGS">FIG. 7</figref> illustrates that an effective refractive index of 0.04 shown in curve <b>712</b> provides significantly greater lateral confinement at smaller waveguide widths than an effective refractive index of 0.004 shown by curve <b>716</b>. The advantage is less pronounced when wider waveguides are used, however wider active areas are typically not preferred due to difficulties with carrier confinement.

0057Surrounding the active region with higher bandgap materials not only provides optical index guiding, the higher bandgap can also improve lateral carrier confinement because the lateral increase in the bandgap energy prevents or at least reduces lateral diffusion of the injected carriers. Carrier confinement becomes exponentially more effective with increasing band offsets, so it is desirable to maximize the band offset by incorporating as much hydrogen in the lateral cladding area as practically possible. The physics of carrier confinement is described in, for example, chapter 3 of Physics of Optoelectronic Devices, by Shun Lien Chuang, John Wiley &amp; Sons, 1995, ISBN 0-471-10939-8. Carrier blocking regions enable more efficient current injection resulting in higher power conversion efficiency. Higher efficiencies enable narrower waveguide regions for single mode devices because fewer carriers are lost to lateral diffusion. In order to further enhance current blocking, nitrogen may be used not only in the active light-emitting layers, but also in adjacent cladding and waveguide layers. Introducing nitrogen into the cladding layers allows hydrogenation to bandgap shift a thicker layer and thus further improving lateral carrier confinement.

0058Although <figref idref="DRAWINGS">FIG. 5</figref> shows a buried index guided laser diode structure, it should be understood that the described index guiding and current confining properties may also be used for other laser types. For example, the described hydrogenation techniques may be used in ridge-waveguide laser diode shown in <figref idref="DRAWINGS">FIG. 8</figref>. In ridge-waveguide laser diodes, hydrogenation can be done by simple diffusion from the etched ridge <b>804</b>. The hydrogenation in this example provides an additional “soft” confinement of the optical mode which reduces e.g. light scattering at the etched mesa edges. Alternate gradual control of hydrogenation will be described in connection with waveguides in <figref idref="DRAWINGS">FIG. 14</figref> and the accompanying description, although the gradual hydrogenation control may also be used to guide the signals in the active region of a laser.

0059Variable Wavelength Lasers and Laser Arrays.

0060As modern networks need to carry more information on optical fibers, Wavelength division multiplexing (WDM) emerged as an important method of increasing the carrying capacity of an optical fiber. In WDM, different wavelengths of light are simultaneously transmitted on a common fiber. Each wavelength carries its own information. The different wavelengths are separated or demultiplexed by devices at a receiving end of the fiber. A typical DMW system uses several semiconductor light output sources, typically semiconductor laser sources that emit at different wavelengths. One example WDM system uses an array of lasers that together are capable of outputting over 100 wavelengths spanning a range from 1525 nm to about 1600 nm. Eventually, WDM systems are expected to encompass the S-band nearing 1300 nm, the wavelength approximately output by InGaAsN systems. However, fabricating the laser array to output the many different wavelength outputs is difficult and thus expensive.

0061Laser wavelength output depends on several factors. Typically a laser tends to lase at the mode frequency with the greatest net gain (net gain=stimulated emission minus optical losses), see, Yariv, A. 1991 Optical Electronics, 4^thedition (Holt, Rinehart and Winston). Thus, selecting an output wavelength involves adjusting the greatest net gain position.

0062<figref idref="DRAWINGS">FIG. 9</figref> shows the various elements that can go into determining the net gain. Net gain is typically determined by the sum of the material gain shown in curve <b>904</b> and a free spectral range curve shown in curve <b>908</b>. Curve <b>904</b> shows the gain in the active material. A peak <b>906</b> of material gain curve <b>904</b> typically corresponds to the bandgap of the active material where the average output energy of a photon is approximately equivalent to the bandgap. Curve <b>908</b> corresponds to the free spectral range of the laser which is determined by the active material laser cavity dimensions. The laser output typically stabilizes at a frequency at the spectral range peak <b>910</b> that is closest to active material peak <b>906</b>. Thus by adjusting the bandgap of the active material using hydrogenation, the output of the laser may be selected.

0063One problem with the illustrated structure is that the peak gain of the active material shifts with temperature. <figref idref="DRAWINGS">FIG. 10</figref> shows an example array of semiconductor lasers including a semiconductor laser cavity <b>1004</b> surrounding active material <b>1008</b>. Furthermore, processing changes make it difficult to repeatedly create a bandgap peak at a desired frequency. Thus a grating may be added to provide feedback to each laser in the laser system. In <figref idref="DRAWINGS">FIG. 10</figref>, each semiconductor laser outputs light into a corresponding grating <b>1012</b>. The grating provides a feedback signal that determines the laser mode selected. This type of laser is called the distributed Bragg reflector (DBR) laser. A discussion of how mode selection is determined relative to grating dimensions and material parameters is given in, for example, pp. 95-101, Diode lasers and Photonic Integrated Circuits, by Larry Coldren and Scott Corzine, John Wiley &amp; Sons, 1995, ISBN 0-471-11875-3. The feedback provided by grating <b>1012</b> is graphically illustrated by curve <b>912</b> of <figref idref="DRAWINGS">FIG. 9</figref> where the laser output wavelength is determined by grating feedback peak <b>916</b>.

0064When fabricating an array of different lasers with different output wavelengths for a WDM system, different gratings are used to produce the different laser outputs. One fabrication challenge is to approximately match active material curve peak <b>906</b> to the grating feedback peak <b>916</b>. Peak mismatches result in inefficiencies, and in the worst case, insufficient total gain to maintain continuous stimulated emission output. Thus, the active material bandgap should be adjusted thereby adjusting active material curve peak <b>906</b> to approximately match grating feedback peak <b>916</b>.

0065One method of adjusting active material curve peak <b>906</b> is to change the growth recipe of each laser device. In particular, the active layer of each laser may differ according to the desired output frequency. The variation may be achieved by varying the epitaxial growth of each active layer as described in U.S. Pat. No. 4,839,899 entitled “Wavelength Tuning of Multiple Quantum Well (MQW) Heterostructure Lasers” by Burnham et al. which is hereby incorporated by reference. However, this multiple recipe approach is difficult to implement in manufacturing settings. In particular, applying different recipes can result in delays and reproducibility problems associated with recipe changes. As used herein, recipe means any sequence of instructions and parameters used to form a particular device. Multiple recipes also increase costs due to the increased length of time required to build the device.

0066Instead of altering each active region during growth, one exemplary embodiment of the invention adjusts the wavelength output of each laser by hydrogenation-induced bandgap tuning (hereinafter HIBS). HIBS adjusts the bandgap of each InGaAsN laser by changing the amount of hydrogen in each laser's active region. Use of HIBS allows multiple lasers designed to output different wavelengths to be grown using the same reactor and the same recipe and even on the same common InGaAsN layer. After growth, the hydrogen content in each active region of each laser is adjusted. As previously described and illustrated in <figref idref="DRAWINGS">FIG. 1</figref>, increasing the hydrogen concentration increases the bandgap of the InGaAsN/GaAs quantum wells thereby shortening the wavelength at which active material peak <b>906</b> will occur. Thus shorter wavelength lasers with higher hydrogen concentrations in the active material will correspond to lasers that output shorter wavelength light and are thus coupled to gratings with shorter periodicities or spacings between adjacent grates.

0067The same hydrogenation and grating shift techniques used DBR lasers may also be used in Distributed Feedback Lasers (DFB). In a DFB laser, the gratings are formed above the active layer instead of beyond it as in DBR lasers. DFB lasers are described in, for example, pp. 102-108, Diode lasers and Photonic Integrated Circuits, by Larry Coldren and Scott Corzine, John Wiley &amp; Sons, 1995, ISBN 0-471-11875-3 which is hereby incorporated by reference.

0068Using the above described methods, an InGaAsN laser may typically be tuned in a range from 1.3 to 1.55 micrometers. The hydrogen concentration determines the output frequency which determines the grating spacing used. For example, if an InGaAsN laser is designed to output a wavelength of 1.55 μm, a typical grating pitch separation of about 502 nm (half wavelength pitch with 114.7 nm of epi and 387.5 nm air gap) is appropriate. The actual grating separation can be scaled according to the material refractive index to produce half wavelength pitch separation if materials other than air are used between grating features. Higher order grating geometries where the pitch is made an integer number of half wavelengths is also possible. Shorter wavelength emitters on the same substrate have grating pitches accordingly scaled shorter to maintain a half wavelength pitch.

0069In one embodiment, the active material is hydrogenated to shift the peak material gain closer to the target wavelength. <figref idref="DRAWINGS">FIG. 1</figref> approximates the amount of hydrogen needed for a desired shift. Increasing hydrogen concentrations increases the bandgap towards that of a nitrogen-free alloy. Bandgap adjustments of up to 50 meV may be obtained at sufficiently high hydrogen dosage for In_0.34Ga_0.66As_1-yN_ywith a nitrogen content y of 0.007. Larger nitrogen concentrations enable even larger bandgap shifts by hydrogenation.

0070Various methods may be used to change the hydrogen content in the active region of a semiconductor laser. One previously described technique uses ion implantation. Another method that is particularly suitable to fabricating multiple semiconductor lasers on a wafer that output different wavelengths is to form a mask over the wafer. <figref idref="DRAWINGS">FIG. 11</figref> shows a set up that may be used to hydrogenate laser active regions of a wafer.

0071In <figref idref="DRAWINGS">FIG. 11</figref>, a wafer <b>1104</b> is placed in a vacuum chamber <b>1108</b>. By providing different level of hydrogenation various devices can be fabricated thereon based on a single common epitaxial layer, such as semiconductor lasers, amplifiers, detectors, waveguides and similar structures. A vacuum pump <b>1112</b> removes reactive and other unnecessary gasses along flow path <b>1116</b>. A hydrogen source <b>1120</b> provides hydrogen. Hydrogen source may typically obtain the hydrogen by many methods, including electrolysis of water or decomposition of ammonia. Also H₂or plasma hydrogenation could be used. Hydrogen is provided along flow path <b>1124</b> into vacuum chamber <b>1108</b>. The wafer is heated to a temperature between 150 and 500 degrees centigrade, and more typically between 250 and 400 degrees centigrade to facilitate diffusion of hydrogen into the wafer. The exposure period varies according to the level of hydrogenation desired.

0072A mask deposited over wafer <b>1104</b> may include apertures that correspond to the active regions of each laser. In one embodiment, a half tone mask with different density of apertures or openings over each active region may be used. Areas with a higher density of apertures will result in higher levels of hydrogen incorporation and undergo a larger bandgap shift. <figref idref="DRAWINGS">FIG. 12</figref> shows an example mask that includes aperture regions <b>1204</b> and <b>1208</b>. The higher density of openings in aperture region <b>1204</b> results in a higher hydrogen density and thus a larger bandgap. Thus the laser that will be associated with aperture region <b>1204</b> will output a shorter wavelength. In alternate embodiments, the apertures may be different sizes with larger apertures being used to simulate a higher density of apertures. In still a third alternative embodiment, instead of apertures, the mask may be made of a material partially permeable to hydrogen such that the rate of hydrogen flow is controlled by the thickness of the mask.

0073Post Growth VCSEL Tuning

0074The described method of tuning a semiconductor light output device, typically an active region of a laser is not limited to edge emitting lasers or even lasers. The tuning technique described may also be used for the tuning of the active region of almost any semiconductor device that outputs light including, but not limited to light emitting diodes (LEDs including superluminescent LEDs), amplifiers and vertical cavity surface emitting lasers (hereinafter VCSEL). An example VCSEL <b>1300</b> shown in <figref idref="DRAWINGS">FIG. 13</figref>. The VCSEL tuning may occur during or after VCSEL device fabrication.

0075A VCSEL typically includes an active region, here an InGaAsN active cavity <b>1304</b>. A bottom set of distributed brag reflectors (DBR mirrors) <b>1308</b> supports a bottom surface of active cavity <b>1304</b>. In the illustrated embodiment, bottom set of DBR mirror stack <b>1308</b> includes 35 pairs of DBR mirrors typically formed from AlGaAs. An upper set of DBR mirror stack <b>1312</b> bounds a top surface of active cavity <b>1304</b>. In the illustrated embodiment, upper DBR mirror stack <b>1312</b> include approximately 25 DBR mirror pairs formed from AlGaAs.

0076The DBR mirror stacks are similar to gratings in that the stack feeds back a particular frequency of laser light equivalent to grating curve <b>912</b> of <figref idref="DRAWINGS">FIG. 9</figref>. The vertical dimensions of active cavity <b>1304</b> determine the free spectral range or internal cavity curve <b>908</b> of <figref idref="DRAWINGS">FIG. 9</figref>. In one embodiment, it is desired that the thickness of the active cavity <b>1304</b> be approximately three times the wavelength of the desired VCSEL light output.

0077Low cost coarse WDM systems typically have wavelength spacings of about 25 nm between channels, while dense WDM systems use wavelength spacings of less than 4 nm. However during actual fabrication it is difficult to form the cavity structure with such precision. Thus, wavelength differences from run to run and wafer to wafer are unavoidable due to layer thickness and material composition variations that occur during epitaxial growth. Emission wavelength differences of up to many tens of nanometers are not unusual, even among devices on the same wafer.

0078In order to compensate for those wavelength differences, the active material in VCSEL active cavity <b>1304</b> may be exposed to hydrogen (or hydrogen removed) to compensate for the variations which occur during processing. In particular, out of specification VCSELs may have hydrogen concentrations adjusted (either increased or decreased) to “tune” the VCSEL to output the desired frequency.

0079The mechanism of how a VCSEL wavelength is changed using hydrogenation was described by <figref idref="DRAWINGS">FIG. 9</figref> and the accompanying description. The VCSEL laser cavity <b>1304</b> may be viewed as a structure similar to active laser cavity <b>1008</b> except that the wave propagation in VCSEL <b>1304</b> is in a vertical direction. Thus, similar to tuning laser cavity <b>1008</b>, the output wavelength of the VCSEL may be tuned by shifting curve <b>908</b> of <figref idref="DRAWINGS">FIG. 9</figref> such that the peak occurs at or near the desired frequency. Thus, if the wavelength output by a VCSEL is too long, the output wavelength may be reduced by increasing the hydrogen concentration in the InGaASN/GaAS quantum well. If the output wavelength is too short, hydrogen may be removed from the active region. As previous described, increased hydrogen concentration increases the active material bandgap resulting in a shorter wavelength output.

0080Various methods may be used to adjust the hydrogen concentration in the VCSEL. In one embodiment tuning is done by exposing the VCSEL to hydrogen. In particular, the VCSEL may be exposed to a hydrogen gas at a temperature of 300 degrees centigrade to facilitate incorporation of hydrogen. Alternately, hydrogen ion implantation may be used to embed hydrogen atoms into the material. In still an alternate embodiment, hydrogen may be incorporated into the VCSEL structure during epitaxial growth, and the adjustment may consist of annealing out excess hydrogen to tune the VCSEL after fabrication.

0081The VCSEL material is grown by Metal Organic Vapor Phase Epitaxy (MOCVD), where hydride precursor gases such as ammonia are typically employed. Moreover, hydrogen is usually the carrier gas of choice for flowing metal organics for building the epitaxial structure. The hydrogen-rich environment in the entire growth process makes it easy to incorporate hydrogen into the material during epitaxial growth. Excess hydrogen can then be driven out of the system by annealing the material at about 500 deg. C. or above during device fabrication.

0082In the case of ion implantation, selective areas of the material can be masked off by photoresist during device fabrication. The unmasked surfaces can then be exposed to standard hydrogen ion implantation accelerated at, typically, between 50 KeV to 350 KeV depending on the penetration depth desired. In practice, the exact dosage and energy levels required is determined experimentally using Secondary Ion Mass Spectroscopy (SIMS) characterization in conjunction with electroluminescence measurements. Monte Carlo computer simulation techniques using software such as the popular SRIM and TRIM packages can also be employed to model the ion implantation process.

0083Although the prior description has focused on tuning a VCSEL, other application of hydrogenation in VCSEL fabrication are also valuable. For example, hydrogenation may be used to adjust the lateral index profile in a VCSEL structure to improve the electrical and optical confinement as previously described for an edge emitter laser. Hydrogenation can also be used to stabilize the polarization of the light output of a VCSEL. In particular, hydrogenation may be used to create refractive index asymmetries in the VCSEL aperture. These asymmetries can be used to control the polarization of the output. An example of controlling polarization by creating asymmetries is discussed in U.S. Pat. No. 6,304,588 entitled “Method and Apparatus for Eliminating Polarization Instability in Laterally-Oxidized VCSELs” by Chua et al which is hereby incorporated by reference in its entirety. In this case, the same techniques may be used, however instead of lateral oxidation, a hydrogenation process is used is used to create the asymmetries.

0084Linking Circuitry:

0085An additional application of Hydrogen induced bandgap and index changes (HIBs) is to create passive optical and integrated optic components on the same wafer as active elements. Various passive and active elements can be formed on the same wafer and even in the same epitaxial layer. This is accomplished by selective area hydrogenation (or dehydrogenation) during device/circuit fabrication.

0086Passive optical elements such as waveguides are often used to interconnect devices. One method of forming such a waveguide is on a semiconductor wafer. To create a waveguide, vertical and lateral optical confinement is needed. The layer structure of the wafer provides vertical optical confinement. Vertical confinement dimensions in such waveguides are usually much smaller (typically &lt;1 μm) than the lateral dimensions (few μm to tens of μm). Thus the index changes used to maintain single mode operation of the waveguide are larger in the vertical than in lateral dimensions. The index profile in vertical direction is given by the layer structure and can therefore be controlled precisely by adjusting appropriate growth parameter (layer thickness and composition). The relatively small index changes in lateral dimensions are usually produced by etching a ridge in the upper cladding layer to produce a lateral effective index change. Etching a ridge involves very well controlled etching. An alternate approach use narrower lateral dimensions by etching a deep ridge through the waveguide core. In order to maintain single mode operation a higher bandgap epitaxial material is usually grown beside the ridge with a subsequent regrowth step. However, the regrowth results in a non-planar morphology of regrown structures.

0087Hydrogen induced bandgap and refractive index shifting of an InGaAsN waveguide layer offers a simple planar processing technique for fabricating passive optical components as well as integrating the passive optical components with active components. As used herein, passive components are broadly defined to include waveguides and similar elements including, but not limited to beam splitter, coupler and taper structures. Passive elements do not undergo electrical stimulation to generate light. Conversely, “active components” are defined as any component that generates light or an electrical signal output from an energy input such as an optical pump or an electrical current. Example active devices include but are not limited to lasers, optical amplifiers, LEDs, electro optical intensity or phase modulators and photo detectors.

0088Many active elements as well as the passive elements can be fabricated from the same layer and even the same epitaxial layer structure. Usually the most complex component (e.g. laser structure) determine the layer structure. Compromises to allow the fabrication of different active elements are sometimes necessary. Hydrogenation processes define the areas used for the different components. In general, hydrogenation is avoided within laser, amplifier and photo diode sections. In other devices such as modulators, a low hydrogenation dose is used to shift the band gap. A medium hydrogenation may be used for most passive elements such as waveguides and a strong hydrogenation around the various components to help confine the optical and electrical signals.

0089Hydrogenation induced band gap and refractive index shifts can also be used to create a desired lateral or modify a given vertical mode profile within a waveguide components. For example, the index shown in curve <b>2304</b> of <figref idref="DRAWINGS">FIG. 23</figref> may be used within a beamsplitter to concentrate the modes into two primary lobes prior to splitting. The index shown in curve <b>2204</b> of <figref idref="DRAWINGS">FIG. 22</figref> may be used to direct optical energy to one side of a waveguide when an asymmetric distribution of light in a waveguide is to be used at the perimeter or “lateral edges” of the active and passive components, a step or other sharply defined hydrogenation profile can be used to create an optical confining barrier. Alternately, a more gradual hydrogenation may be used to create a desired index profile and therefore to design a desired lateral light distribution. Graded index changes may be used to optimize matches between the optical mode profiles in the various different components and waveguides.

0090<figref idref="DRAWINGS">FIG. 14</figref> shows a front cross sectional view of a passive waveguide. In <figref idref="DRAWINGS">FIG. 14</figref>, the waveguide <b>1400</b> is formed on a wafer upon which lasers similar to the laser structure of <figref idref="DRAWINGS">FIG. 5</figref> has been formed. Thus in the example embodiment, laser component layers including cladding layers <b>1404</b>, spacer layers <b>1408</b> and active layers <b>1412</b> are already deposited The waveguide of <figref idref="DRAWINGS">FIG. 14</figref> uses the layers that already exist on the wafer to form a waveguide.

0091In waveguide <b>1400</b>, spacer layer <b>1408</b> and active layer <b>1412</b> can together collectively serve as a set of passive waveguide “core” layers <b>1414</b>. The layer structure provides vertical optical confinement. Adjusting hydrogen concentrations in lateral regions <b>1416</b>, <b>1420</b> creates a lateral variation in the refractive index in nitrogen containing layers, specifically the core layers. By laterally confining a portion of core layer <b>1414</b>, the actual waveguide core <b>1415</b> can be formed. In one exemplary embodiment, because the waveguide core is fabricated from the core layer which is the same layer used to form a laser active region, both the laser active region and the waveguide core <b>1400</b> have approximately the same composition.

0092In <figref idref="DRAWINGS">FIG. 14</figref>, a half tone or gray scale mask <b>1424</b> determines the lateral index profile. When exposed to hydrogen <b>1428</b>, mask <b>1424</b> is toned to control hydrogen diffusion into the waveguide core layers. As used herein, “toned” is broadly defined to mean any of many methods to control the flow of hydrogen through a mask region. In one exemplary embodiment, such toning may be done by making the tone mask thicker in areas in which less hydrogen is desired, and thinner in areas where more hydrogen is needed. An example of such a suitable toning material is photoresist patterned using grayscale photolithography, a well-known micro-machining processing technique. At sufficiently high temperature hydrogen is able to penetrate most materials. However, in practice relative penetration is the most important factor. Thus using a material with varying thickness as a gray scale mask should work for many materials. Examples are SiO₂which has been used for masked hydrogenation of QW lasers (see e.g., Jackson et al, APL 51, 1629 (1987) or Polyimide which has been used for hydrogenated-channel FETs (e.g., see Constant et al, Electron. Lett. 23, 841 (1987).

0093In another exemplary embodiment “toning” may mean the mask material is relatively impervious to hydrogen, but the mask may include small apertures that allow hydrogen to pass through such that areas higher hydrogen concentration are “toned” to allow more hydrogen through by having a higher concentration of apertures as originally shown in <figref idref="DRAWINGS">FIG. 12</figref> or by using larger apertures.

0094The lateral distribution of hydrogen in the core layers around the actual waveguide core determines the lateral index of refraction profile. The lateral index profile determines the shape of the optical mode propagating in the core and thus the optical mode propagating in the waveguide. By controlling the lateral distribution of hydrogen in the core layer, waveguide <b>1400</b> may be made into a single mode waveguide. Single mode waveguides are particularly useful for providing stable conditions (no noise due to light coupling between different modes) for accepting light from or coupling light into other elements including, but not limited to lasers modulators, glass fiber optic cables and light detectors.

0095In the example embodiment of <figref idref="DRAWINGS">FIG. 14</figref>, mask <b>1424</b> gradually and a symmetrically thins with increasing distance from the waveguide core. The thinning mask causes the hydrogen concentration and thus the refractive index to gradually increase with increasing distance from the waveguide core. Although symmetry may be desirable for a standard waveguide, other applications may prefer an asymmetrical mask toning and thus an asymmetrical hydrogen distribution in lateral regions <b>1416</b> and region <b>1420</b>. By making a toning less gradual on one side, a steeper change in index of refraction may be achieved on the one side. A steeper index change results in higher modal confinement on the one side. The higher modal confinement is useful in certain applications. For example, when an asymmetrical beam splitter is needed, tighter modal confinement on one side pulls more light to the one side prior to splitting. In some cases even very asymmetric light distribution are desired. This can for example be achieved by having an index gradient all over the core region and not only an asymmetric profile at the edges. Such a situation is sketched in <figref idref="DRAWINGS">FIG. 22</figref>.

0096As previously described, vertical confinement is usually achieved by the different layer materials. In a laser, the confinement in both the lateral and vertical direction is typically very tight to facilitate electrical pumping current and optical waveguide overlap. However, after entering a passive waveguide, electrical and optical overlap is no longer needed. Thus the waveguide vertical and lateral confinement characteristics may be adjusted to enhance optical coupling between the waveguide and devices coupled to the waveguide. Thus a method of adjusting the vertical confinement factors will be described.

0097<figref idref="DRAWINGS">FIG. 15A</figref> shows a side cross sectional view of a semiconductor DFB or DBR laser <b>1508</b> positioned such that a facet <b>1516</b> is oriented to direct light into waveguide <b>1504</b>. In one embodiment, <figref idref="DRAWINGS">FIG. 15A</figref> may be a side cross sectional view along the core of the waveguide; whereas <figref idref="DRAWINGS">FIG. 14</figref> shows a cross sectional view perpendicular to the waveguide direction of the same waveguide <b>1400</b>. A slight trench <b>1520</b> is etched between the laser <b>1508</b> and waveguide <b>1504</b>. Trench <b>1520</b> facilitates electrical confinement and isolation by preventing current from traveling from contact <b>1524</b> through the passive waveguide <b>1504</b>, thus helping to confine current to semiconductor laser <b>1508</b>.

0098One method of modifying optical vertical confinement is to adjust the refractive index differential between the core layers <b>1540</b><b>1544</b> and the waveguide layers <b>1532</b>, <b>1536</b> in the vertical direction (perpendicular to the waveguide core axis and in the plane of the drawing). Such adjusting can be done by hydrogenation of the waveguide core. <figref idref="DRAWINGS">FIG. 15A</figref> shows gradually increasing the hydrogen concentration along the waveguide core in the direction indicated by arrow <b>1512</b>, the refractive index along the axis of the waveguide core (along a direction parallel to arrow <b>1512</b>) also gradually increases. Because the cladding layers of the waveguide, layers <b>1532</b>, <b>1536</b> do not contain nitrogen, the bandgap of the cladding layer remains approximately constant along the axis of the waveguide. Thus, in the example shown, the gradual increase in hydrogenation along the waveguide core results in a decrease in the index differential between core layers <b>1540</b>, <b>1544</b> and waveguide layers <b>1532</b>, <b>1536</b>. A decreasing index differential results in mode spreading as the wave propagates away from laser <b>1508</b>. This mode spreading improves the vertical optical far field distribution which in turn improves the transverse far field pattern. <figref idref="DRAWINGS">FIG. 15B</figref> shows the refractive index plotted at two points, Point A and Point B of in the structure of <figref idref="DRAWINGS">FIG. 15A</figref>. The plot shows the index of refraction profile gradually decreasing with increasing distance from the semiconductor laser.

0099Thus, by controlling the hydrogen concentration along the waveguide core as shown in <figref idref="DRAWINGS">FIG. 15</figref>, the optical modal profile in the vertical direction can be improved. Likewise, by controlling the hydrogen concentration on either side of the waveguide core as shown in <figref idref="DRAWINGS">FIG. 14</figref>, the optical modal profile in the lateral direction can be controlled. The improvement of the modal profile in both the lateral direction and the vertical direction can be used to improve the transverse far field pattern.

0100The improvement of the transverse far field pattern is particularly helpful in applications where light from a laser <b>1608</b> is being coupled into a glass fiber <b>1604</b> a top view of which is shown in <figref idref="DRAWINGS">FIG. 16</figref>. Different optical modes in the laser and the fiber complicates optical coupling. In particular, the angle at which an optical fiber can receive light, the acceptance angle, is often small, typically less than 20 degrees. By creating an appropriate far field pattern, in particular by using a section of the previously described semiconductor waveguide <b>1612</b>, the input of the fiber can be better matched to the output of the semiconductor waveguide.

0101The top view of the laser and fiber structure of <figref idref="DRAWINGS">FIG. 16</figref> shows section <b>1616</b> a heavily hydrogenated region to create the lateral confinement for an optical signal in a waveguide <b>1612</b> (in one example, a cross section of waveguide <b>1612</b> is the cross section of waveguide <b>1400</b> shown in <figref idref="DRAWINGS">FIG. 14</figref>). The gradual reduction of hydrogenation along the waveguide axis is shown by arrow <b>1620</b> which might be considered analogous to arrow <b>1512</b> of <figref idref="DRAWINGS">FIG. 15</figref>.

0102Traditional fiber optic matching techniques include designing traditional taper structures and fabrication the tapers. (e.g., as described in Y Shani et al. Applied Physics Letters 55, 2389 (1989) or Kasaya K, Mitomi O, Naganuma M, Kondo Y and Noguchi Y 1993 A simple laterally tapered waveguide for low-loss coupling to single-mode fibres IEEE Photon. Technol. Lett. 3 345 and references therein) Taper fabrication uses complicated etching and regrowth operations. By gradually adjusting the hydrogen concentration in the waveguide section <b>1612</b> between the laser <b>1608</b> and a fiber, such as glass fiber <b>1604</b>, many different taper equivalent structures may be formed by a simple planar processing step. In particular, hydrogen induced bandgap changes can be used to taper the semiconductor waveguide and thereby transform the optical mode profile between different components on an opto-electronic integrated circuit (OEIC) or to enable efficient coupling from the OEIC into glass fibers.

0103The waveguides of <figref idref="DRAWINGS">FIG. 14-15</figref> may be used to interconnect and integrate multiple active and passive devices on a single semiconductor wafer. For example, semiconductor lasers with monolithically integrated modulators are commercially available for InGaAsP systems. <figref idref="DRAWINGS">FIG. 17</figref> shows a laser <b>1704</b> and modulator <b>1708</b> section coupled by a hydrogen induced bandgap shifted passive waveguide section. Note that modulator and laser can be fabricated from the same epi-structure. Modulator and amplifier sections can share the same InGaAsN active region; however the amplifier and modulator sections require active regions with slightly different band gap energies in order to reduce the absorption loss in the modulator section. By employing the hydrogen induced bandgap shift in InGaAsN material this can be easily done. Two-section laser diodes, where modulator and amplifier section share the same longitudinal cavity and conventional laser modulator combinations are demonstrated in other materials, (e.g., M. Suzuki, Y. Noda, H. Tanaka, S. Akiba, Y. Kushiro, and H. Isshiki, J. Lightwave Technol. 5, 1277˜1987!. Or T. Tanbun-Ek, Y. K. Chen, J. A. Grenko, E. K. Byrne, J. E. Johnson, R. A. Logan, A. Tate, A. M. Sergent, K. W. Wecht, P. F. Sciortine, and S. N. G. Chu, J. Cryst. Growth 145, 902˜1994 or R. M. Lammert, G. M. Smith, S. Hughes, M. L. Osowski, A. M. Jones, and J. J. Coleman, IEEE Photonics Technol. Lett. 8, 797˜1996). In traditional systems, the bandgap shift between modulator and amplifier section is normally achieved by selective area growth or with a separate regrowth step, which makes the fabrication of devices in these other materials systems quite elaborate and complicated. The preceding example of a laser modulator combination has been provided as one example to demonstrate how interconnect and integrate different active elements (laser with different emission wavelength, optical, light intensity modulator, phase modulators and photo detectors) can be integrated and interconnected with various passive elements such as passive waveguides, beam splitter, (wavelength selective) coupler and taper structure. The example should not be interpreted to limit the types and methods by which these devices may be interconnected.

0104Using the techniques shown in <figref idref="DRAWINGS">FIGS. 14 and 15</figref>, the control of vertical and lateral index profiles in semiconductor material using hydrogenation can be used to fabricate other passive devices such as beamsplitters. <figref idref="DRAWINGS">FIG. 18-22</figref> show exemplary embodiments of various circuits that may be fabricated using the techniques previously described, particularly the techniques described in <figref idref="DRAWINGS">FIGS. 14-15</figref>. Thus all the structures of <figref idref="DRAWINGS">FIG. 18-22</figref> may be fabricated on a single wafer with the interconnects between the lasers, diodes, amplifiers and detectors being fabricated on the same wafer as the devices. This integration greatly facilitates device fabrication.

0105In <figref idref="DRAWINGS">FIG. 18</figref>, two methods are shown of coupling a laser diode to a monitor diode. In a first configuration <b>1804</b>, a beam splitter <b>1812</b> splits the signal from laser diode <b>1808</b>. Some of the light is guided to a monitor diode <b>1816</b> while the remaining light travels along a hydrogenated semiconductor waveguide <b>1818</b> to a taper <b>1820</b> where it couples to other devices or waveguides such as a glass fiber optic cable. In a second configuration, a hydrogenated semiconductor waveguide <b>1830</b> carries signals from a laser diode <b>1828</b> to a monitor device <b>1824</b>. Note that monitor diode, laser and passive waveguide and splitter may be fabricated from the same epilayer.

0106<figref idref="DRAWINGS">FIG. 19</figref> shows coupling a laser diode <b>1904</b> to various amplifiers <b>1908</b> using waveguides and beamsplitters formed by hydrogenating lateral regions on a semiconductor wafer. In the example of <figref idref="DRAWINGS">FIG. 19</figref>, an amplifier is a structure that provides optical gain and in one example has a structure similar to a laser structure except that no optical feedback is provided. In general, an amplifier is any structure that receives an optical input and provides gain to that input.

0107The structure of <figref idref="DRAWINGS">FIG. 20</figref> can also be completely formed on a single wafer. <figref idref="DRAWINGS">FIG. 20</figref> shows an integrated receiver/transmitter module comprising a laser diode, a photodiode, waveguide sections and a wavelength selective waveguide coupler. The laser diode <b>2004</b> outputs light at a first frequency (in this example, 1.55 micrometers) along a semiconductor waveguide <b>2008</b> to a taper for output to external devices such as a glass fiber cable. The coupler directs an incoming light signal (e.g. 1.3 μm) from the glass fiber to photodiode <b>2016</b>. In the illustrated embodiment, the coupler consists of two waveguide sections <b>2008</b> and <b>20012</b> fabricated very close to each other so that the optical modes are coupled via evanescent waves. Coupling length <b>2020</b> is chosen such that light with a wavelength of 1.3 μm is coupled almost completely whereas 1.55 μm light passes the coupler almost unaffected.

0108<figref idref="DRAWINGS">FIG. 21</figref> shows an exemplary embodiment of a complicated circuit, an OEIC (opto-electronic integrated circuit) for optical routing including various beam splitters, light modulators and amplifier sections that may be fabricated on a single wafer. The circuit includes arrays of amplifier/attenuators <b>2104</b>, coupled by various semiconductor waveguides. The purpose of this device is to direct light from a selected entrance to one defined output. For example, in order to connect input <b>2</b> with output <b>8</b>, amplifier/attenuator sections close to input <b>2</b> and <b>8</b> are switched to light amplification while maintaining all other amplifier/attenuator sections in a light attenuation state. Thus the only transparent pass is between gate <b>2</b> and <b>8</b>. By using such a structure, beam splitter internal losses and waveguide attenuation can be (over)compensated in the amplifier sections. The illustrated configuration is particularly suitable for directing one particular input gate to one or more output gates in a multiplexer type of structure.

0109Although the preceding description includes numerous details including hydrogen concentrations, optical circuit configurations, different methods to change hydrogen concentration, and different laser designs. For example, devices such as lasers are described although other components such as light emitting diodes and amplifiers may also use the techniques described herein. Fine details such as the active layer composition as AlGaAsN or InGaAsN are used as examples to facilitate understanding of the device, and to provide example structures. However, these details are not intended nor should they be used to limit the invention as alternate embodiments are understood to be possible without changing the intended invention. Thus the invention should only be limited by the claims, as originally presented and as they may be amended, encompass variations, alternatives, modifications, improvements, equivalents, and substantial equivalents of the embodiments and teachings disclosed herein, including those that are presently unforeseen or unappreciated, and that, for example, may arise from applicants/patentees and others.